Our previous studies have shown that cytomegalovirus (CMV) can infect long- term marrow cultures (LTMC) and cause significant myelosuppression. The mechanism of suppression differed among the various virus isolates used in the study. In some cases direct infection of myeloid precursors could be documented. In other cases various stromal components were infected resulting in alterations of cytokine production. In this application we propose to further dissect the LTMC model, identify the cytokines produced by the major cellular components and determine which of these may be compromised by virus infection. This involves establishing LTMC from normal bone marrow, exposing it to virus, and then assessing the extent of myelosuppression by quantitating changes in the number of myeloid cells produced. Condition media harvested from the LTMC will be evaluated for cytokines using a combination of bioassays, RIA, and ELISA. Cytokine production by distinct cell types will be determined by immunomagnetic selection of cells followed by amplification of cytokine cDNA using polymerase chain reaction (pcr). Standard dual label immune cytochemistry on LTMC will associate cell type with the presence or absence of viral protein, and PCR amplification of viral DNA will identify the presence of viral genome. In order to extrapolate our in vitro LTMC observations to myelosuppression in vivo we will determine if infection of the same cell types occur in vivo. Immunomagnetic techniques will be used to isolate distinct cells from blood and bone marrow obtained from infected patients and isolated cell types will be evaluated for CMV genome by PCR amplification of viral DNA. The presence of CMV in distinct cells types will then be evaluated in relation to clinical outcome. Isolation and identification of cells containing CMV DNA will also be determined for marrow obtained from healthy normal CMV seropositive marrow donors with the eventual goal being to eliminate these cells from the marrow inoculum. The studies proposed in this project will provide an experimental model that should be applicable for studying the hematopoietic suppressive effect of any virus. Initially they will be extended to evaluate the effects of AAV and B19 parvoviruses should these be shown to infect patients posttransplant. Information obtained from this project in conjunction with project 9 should help us to prevent and/or treat viral infections that complicate hematopoietic recovery.